40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  OMC elog  Not logged in ELOG logo
Message ID: 88     Entry time: Mon Apr 1 03:13:41 2013
Author: Koji 
Type: Optics 
Category: General 
Subject: Failure of PZT-glass joints 

[Koji, Jeff, Zach, Lisa]

We glued a test PZT-mirror assembly last week in order to make sure the heat cure of the epoxy does not make any problem
on the glass-PZT joints. The assembly was sent to Bob for the heat treatment. We received the assembly back from Bob on Wednesday.

We noticed that the assembly after the heat cure at 100degC had some voids in the epoxy layer
(looking like the fused silica surface was only 70% "wetted" by the epoxy).
The comparison of the assembly before and after the heat treatment is found in the slideshow at the bottom of the entry.

Initially our main concern was the impact to the control and noise performance.
An unexpected series resonance on the PZT transfer function and unwanted noise creation by the imperfect bonding may terribly ruin the IFO sensitivity.
In reality, after repeated poking by fingers, the PZT-prism joint was detached. This isn't good at all.
Note that there is no sign of degradation on the glass-glass joint.

We investigated the cause of this like:
- Difference of thermal expansion (3ppm/C PZT vs 0.55ppm/C fused silica)
- Insufficient curing of epoxy by UV (but this is the motivation of the heat cure)

Our resolution up to this point is to switch the glue to EP30-2. This means we will go through the heat cure test again.
Unfortunately there is no EP30-2 in stock at Caltech. We asked MIT to send us some packets of EP30-2.

Hardness of the epoxies is another concern. Through the epoxy investigation, we learned from Noliac that the glue for the PZT
should not be too hard (stiff) so as not to constrain the deformation of the PZT. EP30-2 has Shore D Hardness of 75 or more,
while Optocast UV epoxy has 88, and EPOTEK Epoxies, which Noliac suggested for gluing, has ~65. This should also be
confirmed by some measurement.  We will also ask Master Bond if they have information regarding the effect of curing
temperature on the hardness of the epoxy.  EP30-2 can be cured anywhere between RT and 200F (it's service range is up to 300F).
However, the entire breadboard, with the curved mirror sub-assemblies, will need to be baked at 110C to cure the UV Bond epoxy. 
We hope that exposure to relatively higher temps doesn't harden the EP30-2. The EP30-2 data sheet recommends an epoxy
thickness of 80-120 microns which is much thicker than we would like.

We also don't have a way tocontrol the thickness; though we could add glass spheres to the epoxy to control the thickness.
The thickness of the EP30-2 used to bond the metal wire guide prism on the core optics is much thinner at 15-25 microns.

 

ELOG V3.1.3-